Bedford – October 16, 2026 -- Oxipital AI has launched Foreign Material Detection, an AI vision capability that identifies contaminants in real time on high-speed food production lines, with detection accuracy of up to 99% in production environments.
The capability runs on the OX and VX hardware platforms that food producers already use for quality inspection and process automation. It is already operating on food industry production lines and joins quality inspection and vision-guided robotics on the company's unified platform. CEO Mark Chiappetta said customers wanted "fewer vendors, fewer integration headaches" and more confidence in what their inspection systems catch.
Synthetic data trains models before any contamination event occurs
Oxipital AI trains its models on synthetic data rather than waiting for real contamination events. The data simulates hard and soft plastics, wood chips, full and partial PPE, cardboard, metal shards and bone fragments across varied lighting conditions, product types and line speeds. According to the company, models arrive production-ready and begin inspecting on day one.
Precision tuning is meant to keep false positives low, so lines are not halted by natural product variation.
V-CORTX triggers line responses based on contaminant severity
The V-CORTX platform integrates with existing line control systems. A minor anomaly can slow the line, while a confirmed high-severity event stops it without manual intervention. The system also classifies the contaminant type. That lets quality assurance and engineering teams trace whether the source is a worn line component, a supplier issue or a changeover gap.
Every detection event is logged in the V-CORTX analytics dashboard
Each event is recorded with timestamp, classification, imagery, dimensional data and throughput impact. Plant managers and corporate engineering teams get a searchable record that supports FSQA documentation and root cause analysis.
The company notes a limit: vision-based inspection detects only material visible to the camera. Items too small to resolve at operating line speed, or outside the camera's field of view, may go undetected.
Oxipital AI will demonstrate the system live at Pack Expo, booth S-3192, October 18-21.